<< . .

. 57
( : 66)

. . >>

Two more sophisticated approaches are using Monte Carlo simulation and real options, which
are excellent solutions but beyond the scope of this chapter.

CHAPTER 12 Valuing Startups 409
Startups are much riskier ventures than mature businesses. Because of a
lack of sales history and often a lack of market information, a number of
widely varying scenarios are plausible, and the range of outcomes is
much wider and more unpredictable than that of mature businesses.
In a DCF analysis the forecast cash ¬‚ows are supposed to be the
weighted average cash ¬‚ows, with the appraiser having considered the
full range of possible outcomes. However, it is dif¬cult to do this with
such a wide range of possible outcomes. Instead, typically the appraiser,
investment banker, or venture capitalist uses the usually optimistic fore-
cast of the client”perhaps downplayed somewhat”and discounts that
to present value at a very high rate, around 50“75%.
Thus, a more traditional single-scenario DCF analysis to calculate fair
market value is not only more dif¬cult to perform, but it is also far more
subject to criticism by parties with different interests. Short of using
Monte Carlo simulation”a complex approach requiring specialized soft-
ware that is warranted only in a limited number of assignments with
very sophisticated clients”it is virtually impossible to portray the cash
¬‚ows accurately in a single scenario. Instead, the best solution is to use
a multi-scenario approach known as the First Chicago approach. I name
the typical scenarios: very optimistic (the grand slam home run), opti-
mistic (the home run), conservative (the single), and pessimistic (the
According to James Plummer (Plummer 1987), Stanley C. Golder
(Golder 1986) was the originator of the First Chicago approach, named
after First Chicago Ventures, a spinoff of First Chicago Bank™s Equity
Group. In 1980 he founded the venture ¬rm Golder, Thoma, and Cressey.
James Plummer actually gave the name to the First Chicago approach.
Bradley Fowler wrote the original literature on the First Chicago approach
(Fowler 1989, 1990, 1996).

Discounting Cash Flow Is Preferable to Net Income
While discounting forecast cash ¬‚ow is always preferable to discounting
forecast net income, it is even more important to use cash ¬‚ow in valuing
startups than it is in mature ¬rms. This is because cash is far more likely
to run out in a startup than in a mature ¬rm. When that happens, the
¬rm is forced either to take on new investment, which dilutes existing
shareholders™ ownership in the company, or go out of business. In both
cases, using a discounted future net income approach will lead to a se-
rious overvaluation.
When budget is a consideration, it is possible to discount forecast
net income instead of cash ¬‚ow. However, it is critical that the appraiser
at least do some due diligence to ascertain that the subject company will
not run out of cash.

Capital Structure Changes
Startups tend to have somewhat frequent changes in capital structure.
Investment often occurs in several traunches. These changes can involve

PART 5 Special Topics
replacing debt with common or preferred equity and new investment in
equity. This complicates the value calculations because one must be very
careful about whose equity he or she is measuring. Each round of in-
vestment dilutes existing equity, and it is easy to measure the wrong
equity portion if one is not careful.

Venture Capital Rates of Return
Venture capitalists price companies by determining the present value of
cash ¬‚ow or future earnings. One method of valuation is to discount an
optimistic forecast of FMV at the required rate of return. Required rates
of return for VC vary directly with the stage of the company, with star-
tups being the riskiest, hence requiring rates of return of 50“75% (Plum-
mer 1987).
Fowler cites (Fowler 1990) a survey published by Venture Economics
covering 200 companies which indicated that 40% of VC investments lost
money, 30% proceeded sideways or were classi¬ed as ˜˜the living dead,™™
20% returned 2“5 times invested capital, 8% returned 5“10 times, and 2%
returned greater than 10 times the investment. In a follow-up article
(Fowler 1996) he refers to comments made by Professor Stewart Myers
of MIT in his November 1995 address to the American Society of Ap-
praisers con¬rming that 70“80% of VC investments are failures, whereas
20“30% are big winners. In addition, Professor Myers observed that the
overall IRR for successful VC partnerships was approximately 25%.4
The 25% rate of return is consistent with a more recent Wall Street
Journal article (Pacelle 1999) which cites Venture Economics as a source
that venture capital ¬rms returned an average 27.4% over the past 5 years,
although they returned only 15.1% over the past 20 years. From this, we
can calculate the ¬rst 15 years™ (roughly 1979“1993) compound average
return as 11.27%.5 That is a very low return for VC ¬rms. It is comparable
to NYSE decile #1 ¬rm long-run returns. I would attribute that low return
to two factors. That period:
1. Was the infancy of the VC industry, and the early entrants faced
a steep learning curve.
2. Included two severe recessions.
It is not reasonable to expect VC investors to be happy with a 15% return
long run. The ¬ve-year average of 27.4% is more in line with the risk
As to batting averages, a reasonable synthesis of this information is
that 2% of VC investments are grand slams, 8% are home runs, 20% are
moderately successful, and 70% are worthless or close to it.

He also mentioned that the average VC project return was 1%. He said the difference in returns
is due to the skewness in the distribution that comes from the venture capitalists quickly
identifying and pulling the plug on the losers, i.e., they do not continue to fund the bad
projects. Thus, the bad projects have the least investment.
r15)(1.274)5 1.15120, which solves to r15
The equation is: (1 11.27%.

CHAPTER 12 Valuing Startups 411
Table 12-1: Example of the First Chicago Approach
In Table 12-1 we use these percentages for weighting the four different
scenarios, very optimistic, optimistic, conservative, and pessimistic, re-
Initially we perform discounted cash ¬‚ow calculations to determine
the conditional FMV of the subject company under the different scenarios.
Typical venture capital rates of return include the discount for lack of
marketability (DLOM) and discount for lack of control (DLOC). This
tends to obscure the discount rate, DLOM, and DLOC. The appropriate
discount rate using the First Chicago approach begins with the average
success rate of approximately 25% reported by Professor Myers.
The 25%, however, is a geometric average rate of return. We should
estimate an increment to add in order to estimate the arithmetic rate of
return.6 In Table 5-4 we show arithmetic and geometric mean rates of
return from log size model regressions of the 1938“1997 New York Stock
Exchange data for different size ¬rms.
For a ¬rm of $1 million FMV, the regression forecast arithmetic and
geometric returns, rounded to the nearest percent, are 25% and 18%, re-
spectively, for a differential of 7%. For a ¬rm of $25 million FMV, the
regression forecast arithmetic and geometric returns, rounded to the near-
est percent, are 21% and 16%, respectively, for a differential of 5%. We
can add the size-based differential to estimate the arithmetic average rate
of return to use for our discount rate. For most size ranges the result
comes to approximately 30%.7
Column B of Table 12-1 lists the conditional FMVs obtained from
discounted cash ¬‚ow analyses using different sets of assumptions. In the
very optimistic scenario we forecast outstanding performance of the com-
pany, with a resulting FMV of $130,000,000 (B6). Cells B7 and B8 display
the FMVs arising from optimistic and conservative forecasts, respectively.
In the pessimistic scenario we assume the company fails completely, re-
sulting in zero value. When valuing a general partnership interest, which

T A B L E 12-1

First Chicago Method


5 Conditional FMV [1] Probability [2] Wtd FMV

6 Very optimistic scenario $130,000,000 2% $2,600,000
7 Optimistic scenario 50,000,000 8% 4,000,000
8 Conservative scenario 10,000,000 20% 2,000,000
9 Pessimistic scenario [1] 0 70% ”
10 Weighted average FMV 100% $8,600,000

[1] Individual discounted cash ¬‚ow analyses are the source for the numbers in this column
[2] Based on the VC rates discussed in the chapter

I con¬rmed this in a telephone conversation with Professor Myers.
Fowler™s article did not address this adjustment.

PART 5 Special Topics
has unlimited liability, the appraiser should consider the possibility of
negative value.
Column C lists the probability associated with each scenario. These
are derived directly from the empirical probabilities of VC success dis-
cussed above. We calculate the weighted FMV in column D by multiply-
ing the conditional FMV in column B by its associated probability in
column C and summing the results. Thus, in this example the weighted
average FMV is $8,600,000 (D10).

Advantages of the First Chicago Approach
The major advantages of the First Chicago approach are:
1. It reduces the uncertainty associated with a single FMV by
allowing for several scenarios representing differing levels of
success of the company.
2. It breaks down the huge range of potential outcomes into ˜˜bite-
size™™ chunks, i.e., the individual scenarios, that are credible and
plausible when performed carefully.
3. It makes the appraiser™s probability distribution of outcomes
explicit. In doing so, it has two additional advantages: (a) If the
client agrees with the conditional FMVs of each scenario but for
some reason feels the probabilities are not representative of the
subject company™s chances, it is an easy exercise for the client to
weight the probabilities differently and adjust the valuation him
or herself. This is particularly important when the assignment is
to provide existing shareholders with information to negotiate
with funding sources. If both sides accept the scenario
valuations, it is usually easy for them to come to terms by
agreeing on the probabilities of the outcomes, which they can
easily do without the appraiser; and (b) it protects the appraiser.
When the appraiser shows a ¬nal weighting of the conditional
FMVs multiplied by their probabilities to calculate the FMV and
the appraiser shows the probability of total failure as, say, 70%,
it can protect the appraiser from a disgruntled investor in the
event the company fails. The appraiser has clearly
communicated the high probability of investors losing all their
money, despite the fact that the FMV may be very high”and,
we hope, is”due to the large values in the upper 30% of
probable outcomes.
Therefore, the First Chicago approach is normally the preferred
method of valuation of startups. It is also useful in valuing existing ¬rms
that are facing radically different outcomes that are hard to forecast. For
example, I used it recently to assist warring shareholders who wanted
one side to buy out the other in a four-year-old company. The ¬rm was
pro¬table and had grown rapidly, but there were several major uncer-
tainties that were impossible to credibly consider with accuracy in a single
DCF scenario. The uncertainties were as follows:
1. There was much customer turnover in the prior year, despite
healthy growth.

CHAPTER 12 Valuing Startups 413
2. If one of the shareholders left, sales might suffer greatly for two
or three years and even endanger the company.
3. There were regulatory issues that could have had a dramatic
impact on the company.
4. Pro¬t margins were highly variable in the past four years and
could have been affected by regulation.
Collectively, these uncertainties made a single scenario forecast of
sales growth and pro¬tability very dif¬cult. Despite considerable parti-
sanship by the shareholders, who often actively lobbied for changes in
the DCF analyses, the First Chicago approach enabled us to credibly
model the different paths the Company could take and quantify the val-
uation implications of that. Ultimately, we presented them with the val-
uation of the different scenarios and our estimates of the probabilities,
and the weighted average of the product of the two constituted our es-
timate of FMV. We also explained that they could change their subjective
weighting of probabilities of outcomes, thus changing the FMV. Ulti-
mately, they worked out an arrangement without any further need for
our help.

Discounts for Lack of Marketability and Control
Finally, venture capitalists typically have more control and possibly mar-
ketability than most other investors. When valuing the interests of other
investors, the appraiser must add the incremental discounts for lack of
control and marketability that apply to the speci¬c interests, i.e., an arm™s-
length investor would typically require a higher rate of return on smaller
interests than the 30% that the VC expects.

In this approach the appraiser estimates net earnings at cash-out time,
often at Year 5 or 6. He or she then estimates a P/E multiple and mul-
tiplies the two to estimate the cash out.
In Table 12-2 we use Fowler™s (Fowler 1989) numbers, with minor
changes in the presentation. Fowler assumed Year 5 net income of
$1,936,167 and multiplied it by a P/E multiple of 12 to calculate the Year
5 cash out at $23.2 million (B5), rounded.

T A B L E 12-2

VC Pricing Approach [1]


5 Assumed cash out-5 yrs @ 12 earnings $23,200,000
6 Present value factor-5 years @ 45% ROI 0.1560
7 Present value-rounded $ 3,619,000

[1] Source: Bradley Fowler, What Do Venture Capital ˜˜Pricing Methods Tell About Valuation of Closely Held Firms?™™ Business Valua-
tion Review, June 1989, page 77.

PART 5 Special Topics
He then used a 45% rate of return to discount cash ¬‚ows, based on
industry statistics he presented in the article, which we repeat below in
the next section. The present value factor at 45% for ¬ve years is 0.156,
and the present value of the Company is then $3,619,000 (B7, rounded).

Venture Capital Rates of Return
Fowler (1989) cited rates of return from two different studies. Plummer
(1990) found that the required rates of return (ROR), which included dis-
counts for lack of control (DLOC) and discounts for lack of marketability
(DLOM), were:

Stage of Development of Co. Required Rate of Return

Seed capital stage 50“75%
1st stage 40%“60%
2nd stage 35%“50%
3rd stage 30%“50%
4th stage 30%“40%

Morris (1988, p. 55) writes that VCs are looking for the following
rates of return:

Stage of Development of Co. Required Rate of Return

Seed capital stage 50%
2nd stage 30“40%

Summary of the VC Approach
The VC approach is a valid valuation approach, though certainly less
analytically precise than the First Chicago approach. Nevertheless, it is
used by venture capitalists, and it serves as a ˜˜quick-and-dirty™™ valuation
method, on the one hand, and as a useful alternative approach, on the
This concludes Part 1 of this chapter. Part 2 is a complex decision
tree analysis combined with multiscenario valuation.

Early-stage technology-based companies often ¬nd themselves in ¬nan-
cial hot water. They incur large expenses for years during the develop-
ment of a new product. Consequently, they run short of funds and often
require the infusion of venture capital, which may or may not occur. In
the following example”which is based on an actual assignment, with
names and numbers changed”the Subject Company has several possible
events that can impact the probability of obtaining venture capital as well
as surviving as a ¬rm without venture capital, i.e., bootstrapping to

CHAPTER 12 Valuing Startups 415
The Company and its former parent (˜˜the parent™™) share a nearly iden-
tical set of shareholders”well over 100. The president is the major share-
holder of the ¬rm, with effective, but not absolute control. The parent had
lent the Company $1 million to get started as a spinoff, but the debt
would be coming due in four years, and the Company has no way of
paying it off.
The parent proposed the following restructuring of the debt:
1. The parent would convert the debt into $400,000 of convertible
preferred stock”and part of the valuation exercise was to
determine how many shares of preferred stock that would be.
There would be no preferred dividends, but the parent would
have a liquidation preference.
2. The president would have to relinquish a certain number of his
shares in the parent back to the parent, which had a ready
buyer for the shares.
In return for relinquishing his shares to the parent, the president
wants the Company to issue 1.3 million new shares to him. The board of
directors wants an independent appraisal to determine whether the trans-
action is favorable to the other shareholders. This example, however, is
typical of the types of decisions faced by startup ¬rms in their quest for
adequate funding. More importantly, the statistical approach we use in
this valuation is applicable to the valuation of many startups, regardless
of industry.

Key Events
The company president, Mr. Smith, has identi¬ed a sequence of four key
events that could occur, and each one of them increases the Company™s
ability to obtain venture capital ¬nancing as well as to successfully boot-
strap the ¬rm without VC ¬nancing. The events are sequentially depen-
dent, i.e., event #1 is necessary, but not suf¬cient for event #2 to occur.
Events #1, #2, and #3 must occur in order for #4 to occur. These events
1. Event #1: The Company sells its product to company #1. The
conditional probability of this event occurring is 75% (Table
12-3, cell B11).
2. Event #2: The Company sells its product to company #2. The
conditional probability of this happening, assuming event #1
occurs, is 90% (B12).
3. Event #3: The Company sells its product to company #3, which
has a 60% (B13) conditional probability, i.e., assuming event #2
4. Event #4: The Company sells its product to company #4. If the
Company sells its product to company #3, then it has an 80%
(B14) probability of selling it to company #4.
While these four events are all potential sales, the statistical process
involved in this analysis is generic. The four events could just as easily

PART 5 Special Topics
be a mixture of technology milestones, rounds of ¬nancing, regulatory,
sales, and other events.

Decision Trees and Spreadsheet Calculations
Our analysis begins as decision trees, which appear in Figures 12-1 and
12-2. However, careful analysis leads to our being able to generalize the
decision tree calculations mathematically and transform them into ex-

<< . .

. 57
( : 66)

. . >>